16.24. taler-auditor.conf(5)¶
16.24.1. Name¶
taler-auditor.conf - Taler auditor configuration file
16.24.2. Description¶
The configuration file is line-oriented. Blank lines and whitespace at the
beginning and end of a line are ignored. Comments start with #
or %
in the first column (after any beginning-of-line whitespace) and go to the end
of the line.
The file is split into sections. Every section begins with [SECTIONNAME]
and contains a number of options of the form OPTION=VALUE
. There may be
whitespace around the =
(equal sign). Section names and options are
case-insensitive.
The values, however, are case-sensitive. In particular, boolean values are
one of YES
or NO
. Values can include whitespace by surrounding the
entire value with "
(double quote). Note, however, that there are no
escape characters in such strings; all characters between the double quotes
(including other double quotes) are taken verbatim.
Values that represent a time duration are represented as a series
of one or more NUMBER UNIT
pairs, e.g. 60 s
, 4 weeks 1 day
,
5 years 2 minutes
.
Values that represent an amount are in the usual amount syntax:
CURRENCY:VALUE.FRACTION
, e.g. EUR:1.50
.
The FRACTION
portion may extend up to 8 places.
Values that represent filenames can begin with a /bin/sh
-like variable
reference. This can be simple, such as $TMPDIR/foo
, or complex, such as
${TMPDIR:-${TMP:-/tmp}}/foo
. The variables are expanded either using
key-values from the [PATHS]
section (see below) or from the environment
(getenv()
). The values from [PATHS]
take precedence over those from
the environment. If the variable name is found in neither [PATHS]
nor the
environment, a warning is printed and the value is left unchanged. Variables (including those from the environment) are expanded recursively, so if FOO=$BAR
and BAR=buzz
then the result is FOO=buzz
. Recursion is bounded to at most 128 levels to avoid undefined behavior for mutually recursive expansions like if BAR=$FOO
in the example above.
The [PATHS]
section is special in that it contains paths that can be
referenced using $
in other configuration values that specify
filenames. Note that configuration options that are not specifically
retrieved by the application as filenames will not see “$”-expressions
expanded. To expand $
-expressions when using taler-config
, you must pass
the -f
command-line option.
The system automatically pre-populates the [PATHS]
section with a few values
at run-time (in addition to the values that are in the actual configuration
file and automatically overwriting those values if they are present).
These automatically generated values refer to installation properties
from GNU autoconf. The
values are usually dependent on an INSTALL_PREFIX
which is determined by
the --prefix
option given to configure. The canonical values are:
LIBEXECDIR = $INSTALL_PREFIX/taler/libexec/
DOCDIR = $INSTALL_PREFIX/share/doc/taler/
ICONDIR = $INSTALL_PREFIX/share/icons/
LOCALEDIR = $INSTALL_PREFIX/share/locale/
PREFIX = $INSTALL_PREFIX/
BINDIR = $INSTALL_PREFIX/bin/
LIBDIR = $INSTALL_PREFIX/lib/taler/
DATADIR = $INSTALL_PREFIX/share/taler/
Note that on some platforms, the given paths may differ depending on how the system was compiled or installed, the above are just the canonical locations of the various resources. These automatically generated values are never written to disk.
Files containing default values for many of the options described below
are installed under $TALER_AUDITOR_PREFIX/share/taler-auditor/config.d/
.
The configuration file given with -c to Taler binaries
overrides these defaults.
A configuration file may include another, by using the @INLINE@
directive,
for example, in main.conf
, you could write @INLINE@ sub.conf
to
include the entirety of sub.conf
at that point in main.conf
.
Be extra careful when using taler-auditor-config -V VALUE
to change configuration
values: it will destroy all uses of @INLINE@
and furthermore remove all
comments from the configuration file!
16.24.2.1. GLOBAL OPTIONS¶
The “[PATHS]” section is special in that it contains paths that can be referenced using “$” in other configuration values that specify filenames. For Taler, it commonly contains the following paths:
- TALER_HOME
Home directory of the user, usually “${HOME}”. Can be overwritten by testcases by setting ${TALER_TEST_HOME}.
- TALER_DATA_HOME
Where should Taler store its long-term data. Usually “${TALER_HOME}/.local/share/taler-auditor/”.
- TALER_CONFIG_HOME
Where is the Taler configuration kept. Usually “${TALER_HOME}/.config/taler-auditor/”.
- TALER_CACHE_HOME
Where should Taler store cached data. Usually “${TALER_HOME}/.cache/taler-auditor/”.
- TALER_RUNTIME_DIR
Where should Taler store system runtime data (like UNIX domain sockets). Usually “${TMP}/taler-auditor-runtime”.
16.24.2.2. CURRENCY SPECIFICATIONS¶
Sections with a name of the form “[currency-$NAME]” (where “$NAME” could be any unique string) are used to specify details about how currencies should be handled (and in particularly rendered) by the user interface. A detailed motivation for this section can be found in DD51. Different components can have different rules for the same currency. For example, a bank or merchant may decide to render Euros or Dollars with always exactly two fractional decimals, while an Exchange for the same currency may support additional decimals. The required options in each currency specification section are:
- ENABLED
Set to YES or NO. If set to NO, the currency specification section is ignored. Can be used to disable currencies or select alternative sections for the same CODE with different choices.
- CODE
Code name for the currency. Can be at most 11 characters, only the letters A-Z are allowed. Primary way to identify the currency in the protocol.
- NAME
Long human-readable name for the currency. No restrictions, but should match the official name in English.
- FRACTIONAL_INPUT_DIGITS
Number of fractional digits that users are allowed to enter manually in the user interface.
- FRACTIONAL_NORMAL_DIGITS
Number of fractional digits that will be rendered normally (in terms of size and placement). Digits shown beyond this number will typically be rendered smaller and raised (if possible).
- FRACTIONAL_TRAILING_ZERO_DIGITS
Number of fractional digits to pad rendered amounts with even if these digits are all zero. For example, use 2 to render 1 USD as $1.00.
- ALT_UNIT_NAMES
JSON map determining how to encode very large or very tiny amounts in this currency. Maps a base10 logarithm to the respective currency symbol. Must include at least an entry for 0 (currency unit). For example, use {“0”:”€”} for Euros or “{“0”:”$”} for Dollars. You could additionally use {“0”:”€”,”3”:”k€”} to render 3000 EUR as 3k€. For BTC a typical map would be {“0”:”BTC”,”-3”:”mBTC”}, informing the UI to render small amounts in milli-Bitcoin (mBTC).
16.24.2.3. EXCHANGE DATABASE OPTIONS¶
The following options must be in the section “[exchangedb]”.
- IDLE_RESERVE_EXPIRATION_TIME
After which time period should reserves be closed if they are idle?
- LEGAL_RESERVE_EXPIRATION_TIME
After what time do we forget about (drained) reserves during garbage collection?
- AGGREGATOR_SHIFT
Delay between a deposit being eligible for aggregation and the aggregator actually triggering.
- DEFAULT_PURSE_LIMIT
Number of concurrent purses that a reserve may have active if it is paid to be opened for a year.
- MAX_AML_PROGRAM_RUNTIME
Maximum time an AML program is allowed to run. (Optional for taler-auditor.)
16.24.2.4. EXCHANGE POSTGRES BACKEND DATABASE OPTIONS¶
The following options must be in section “[exchangedb-postgres]” if the “postgres” plugin was selected for the database.
- CONFIG
How to access the database, e.g. “postgres:///taler-exchange” to use the “taler-exchange” database. Testcases use “talercheck”.
16.24.2.5. EXCHANGE ACCOUNT OPTIONS¶
An exchange (or merchant) can have multiple bank accounts. The following
options are for sections named “[exchange-account-SOMETHING]”. The SOMETHING
is
arbitrary and should be chosen to uniquely identify the bank account for
the operator. These options are used by the taler-exchange-aggregator, taler-exchange-closer, taler-exchange-transfer and taler-exchange-wirewatch tools.
- PAYTO_URI
Specifies the full payto://-URL of the account. The general format is
payto://$METHOD/$DETAILS?receiver-name=NAME
. Examples:payto://x-taler-bank/localhost:8899/Exchange?receiver-name=Exchange
orpayto://iban/GENODEF1SLR/DE67830654080004822650/?receiver-name=Exchange
orpayto://iban/DE67830654080004822650/?receiver-name=Exchange
(providing the BIC is optional).- ENABLE_DEBIT
Must be set to
YES
for the accounts that the taler-exchange-aggregator and taler-exchange-closer should debit.- ENABLE_CREDIT
Must be set to
YES
for the accounts that the taler-exchange-wirewatch should check for credits. It is yet uncertain if the merchant implementation may check this flag as well.
Additionally, for each enabled account there MUST be another matching section named “[exchange-accountcredentials-SOMETHING]”. This section SHOULD be in a secret/
configuration file that is only readable for the taler-exchange-wirewatch and taler-exchange-transfer processes. It contains the credentials to access the bank account:
- WIRE_GATEWAY_URL
URL of the wire gateway. Typically of the form
https://$HOSTNAME[:$PORT]/taler-wire-gateway/$USERNAME/
where $HOSTNAME is the hostname of the system running the bank (such as the Taler Python bank or the Nexus) and$USERNAME
is the username of the exchange’s bank account (usually matching theUSERNAME
option used for authentication). Example:https://bank.demo.taler.net/taler-wire-gateway/Exchange/
.- WIRE_GATEWAY_AUTH_METHOD
This option determines how the exchange (auditor/wirewatch/aggregator) authenticates with the wire gateway. Choices are
basic
,bearer
andnone
.- USERNAME
User name for
basic
authentication with the wire gateway.- PASSWORD
Password for
basic
authentication with the wire gateway.- TOKEN
Token for
bearer
authentication with the wire gateway.
16.24.2.6. EXCHANGE COIN OPTIONS¶
The following options must be in sections starting with "[coin_]"
and are
largely used by taler-exchange-httpd to determine the meta data for the
denomination keys. Some of the options are used by the
taler-exchange-secmod-rsa to determine which RSA keys to create (and of
what key length). Note that the section names must match, so this part of the
configuration MUST be shared between the RSA helper and the exchange.
Configuration values MUST NOT be changed in a running setup. Instead, if
parameters for a denomination type are to change, a fresh section name should
be introduced (and the existing section should be deleted).
- VALUE
Value of the coin, e.g. “EUR:1.50” for 1 Euro and 50 Cents (per coin).
- DURATION_WITHDRAW
How long should the same key be used for clients to withdraw coins of this value?
- DURATION_SPEND
How long do clients have to spend these coins?
- DURATION_LEGAL
How long does the exchange have to keep records for this denomination?
- FEE_WITHDRAW
What fee is charged for withdrawal?
- FEE_DEPOSIT
What fee is charged for depositing?
- FEE_REFRESH
What fee is charged for refreshing?
- FEE_REFUND
What fee is charged for refunds? When a coin is refunded, the deposit fee is returned. Instead, the refund fee is charged to the customer.
- CIPHER
What cryptosystem should be used? Must be set to either “CS” or “RSA”. The respective crypto-helper will then generate the keys for this denomination.
- RSA_KEYSIZE
What is the RSA keysize modulos (in bits)? Only used if “CIPHER=RSA”.
- AGE_RESTRICTED
Setting this option to
YES
marks the denomination as age restricted (default isNO
). For this option to be accepted the extension for age restriction MUST be enabled.
16.24.2.7. EXCHANGE OPTIONS¶
The following options are from the “[exchange]” section and used by the auditor.
- CURRENCY
Name of the currency, e.g. “EUR” for Euro.
- CURRENCY_ROUND_UNIT
Smallest amount in this currency that can be transferred using the underlying RTGS. For example: “EUR:0.01” or “JPY:1”.
- DB
Plugin to use for the database, e.g. “postgres”.
16.24.2.8. AUDITOR OPTIONS¶
The following options must be in section “[auditor]” for the Taler auditor.
- DB
Plugin to use for the database, e.g. “postgres”
- AUDITOR_PRIV_FILE
Name of the file containing the auditor’s private key.
- PUBLIC_KEY
Crockford Base32 encoded auditor public key. Used by (online) auditor processes that do not have access to the (offline) auditor private key file.
- BASE_URL
Base URL of the auditor, e.g. “https://auditor.demo.taler.net/”
- SERVE
Should the HTTP server listen on a UNIX domain socket (set option to “unix”) or on a TCP socket (set option to “tcp”)?
- UNIXPATH
Path to listen on if we “SERVE” is set to “unix”.
- UNIXPATH_MODE
Access permission mask to use for the “UNIXPATH”.
- PORT
Port on which the HTTP server listens, e.g. 8080.
- BIND_TO
Hostname to which the merchant HTTP server should be bound to, e.g. “localhost”.
16.24.2.9. AUDITOR POSTGRES BACKEND DATABASE OPTIONS¶
The following options must be in section “[auditordb-postgres]” if the “postgres” plugin was selected for the database.
- CONFIG
How to access the database, e.g. “postgres:///taler” to use the “taler” database. Testcases use “talercheck”.
16.24.3. SEE ALSO¶
taler-auditor-dbinit(1), taler-auditor-httpd(1), taler-auditor-offline(1).
16.24.4. BUGS¶
Report bugs by using https://bugs.taler.net/ or by sending electronic mail to <taler@gnu.org>.